(Updated: 2020-07-20 11:48:30 EDT)
So, I have been watching the epidemic in China and Korea since the beginning, but it is still not easy to settle into this reality here in the US.
First of all, everyone, STAY HEALTHY and see what YOU can do!
Second, there are many excellent news articles and information sources, but I’m sharing some more questions (see left panel) and answers for those who are curious about a specific situation in Maryland, US - especially including local data. Maryland is my beautiful adopted home state of over 20 years. The State is doing its best with strong leadership, but this virus arrived here when the country was not ready and is spreading fast unfortunately…
See footnote for further information about data sources. More questions and answers will be added, as more local data become available.
(Sorry, this is not mobile device friendly, and is best viewed on your regular monitor.)
With the reopening, we expect to see - and have seen - increases in new cases. We hope it remains within a manageable level that the health systems can respond effectively.
Nonetheless, where and among whom have the new cases increased most?
Solid line is 7-day rolling average of daily new cases per 100,000 population. Orange segment of the line is the first two weeks after reopening, and red segment is beyond that period.
Among those with COVID-19, the proportions of younger population have increased. Also, the proportions of minority groups, especially Hispanics, have increased. Only 10% of population is Hispanic (as of 2018), but one third of confirmed cases belong to Hipanic population.
It is important to know age composition within each race as well. As of early June, only three states in the US publish such data, and Maryland is not one of them unfortunately.
NOTE on race data: composition excluding cases with unknown race. Hispanic was not reported as a separate category until April 15th.Yes, at least partially. There is a very strong age-pattern in COVID-19 mortality (i.e., the older, the higher risk of dying), and having proportionately more younger people with COVID-19 would lower overall, all-age, case fatality rate, which has declined slightly (See Question 5). In addition, importantly, a recent study showed how hospital care might have improved and saved lives more effectively over time, even just over several weeks.
At the same time, minority populations have lower health insurance coverage and receive lower quality of care than White population in the US. So, having proportionately more minorities means access to quality health care may decrease among people with confirmed cases.
Answers to Question 5 show latest case fatality rate (i.e., deaths per 100 COVID-19 cases) by county and how it has changed in the State. But, how about across different age groups? Case fatality rate has increased continuously in older age groups. The rate of increase is highest among people who are 80 and older.
(Note: case fatality rate including only lab-confirmed COVID-19 deaths and cases. See Question 5 for comparison between case fatality rates with vs. without probable deaths/cases.)The Maryland Department of Health started publishing data by race (i.e., the number of cases and deaths by race) on April 9th, following an upsetting report about racial disparity in COVID mortality in the US: higher morality in states with higher proportions of black population. Further data - specifically disaggregated by individual people’s race, beyond the state-level analysis - are crucial to monitor and understand the disparity. Also, though initially unavailable, a separate category for Hispanic population is published on April 15. As a resident of Maryland, I am very proud of the state’s rapid action to publish race data!
Now with Hispanic population disaggregated from “other”, the pattern of incidence and mortality by race/ethnicity can be examined better. In terms of rates (important to compare across races with different population sizes), the incidence rate is substantially higher among Hispanic, followed by African American population (blue bars). And, incidence rate among Hispanic population has increased most rapidly - see the second figure below.
However, case fatality rate is highest among White and Asian Americans (orange bars). This implies that the disproportionately higher number of deaths among African American population in Maryland is because of the higher rate of infection, not because of higher risk of dying among those who are infected. At the same time, it is notable that, though the infection rate is lower, mortality risk is higher among Asian Americans in Maryland.
To understand reasons behind this, we will need to learn more about characteristics of patients by race (e.g., Do Asian Americans with COVID tend to be older and/or have existing conditions in Maryland? Are Hispanic Marylanders with COVID younger than their counterparts?) and any differences in access to health care by race among COVID patients. Also, what can we do to reduce the higher infection rate among African Americans and Hispanic population in Maryland? Finally and importantly, if and when we have better data on race (i.e., less cases with missing race information), the findings on racial disparity may well change (see below note on race data in Maryland).
Hover over each figure to see values and more options.
(Note: case fatality rate including only lab-confirmed COVID-19 deaths and cases. See Question 5 for comparison between case fatality rates with vs. without probable deaths/cases.) (Note: case fatality rate including only lab-confirmed COVID-19 deaths and cases. See Question 5 for comparison between case fatality rates with vs. without probable deaths/cases.)Important note on race data in Maryland:
1. 17 % of cases do not have race information. This is likely because private labs are not required to report race. All data shown here is only based among cases and deaths with known race.
2. In Maryland, “Other” population includes: American Indian and Alaska Native, Native Hawaiian and Other Pacific Islander, and ‘Two or more races’ - accounting for about 3% of total population in the state. Figures do not include “Other” races, given possibility that Hispanic population might have been included in this category initially.
As of 2020-07-20 10:00AM, according to Maryland Department of Health, 3252 COVID-19 deaths have occurred, that had been laboratory-confirmed. In addition, MD Health Department has published the number of probably COVID deaths since April 15. There have been 130 probable COVID-19 deaths, for which death certificate lists COVID-19 as the cause of death but not yet confirmed by a laboratory test. This means case fatality rate can be calculated with vs. without the probable deaths (which are thus probable cases as well). Case fatality rate is:
Given relatively small differences across groups (i.e., by age, sex, and race, results now shown), only laboratory-confirmed COVID-19 deaths and cases are used throughout this report. I will keep monitoring how the two approaches produce different/similar results, and update as needed.
Below figure shows mortality by county - in terms of both absolute number (red bars) and case fatality rate (orange bars) as of 2020-07-20.
Below chart shows the mortality trends, since March 18 when the first COVID-19 death was reported in Maryland. Globally in countries severely affected by the epidemic before US, case fatality rates increased rapidly in the beginning. The rates then stabilized in some of the countries, depending on health systems’ response and characteristics of patient population.
This is a very important question, since the magnitude of testing over time is critical information to understand the epidemic. I got trend data on testing in Maryland from COVID-19 Case Map Dashboard by Maryland Department of Health also COVID Tracking Project for earlier data. Still, data on the number of new tests are not available between 3/12 and 3/28.
As of 2020-07-20 10:00AM, a total of 749279 tests have been conducted. There are about 6 million people in Maryland, and this means 123.9 tests have been conducted in every 1000 people.
However, a high rate of positive test, 3.8 % (average over the latest 7-day data), indicates that testing is still limited to those with symptoms primarily, not based on effective contact tracing. Ideally, the positive test rate should be below 5%.
Note 1: On May 28th, Maryland Health Department started publishing “total testing volume”“. However, it is unclear what a unit of testing is, since the volume is about 15% higher than the sum of”Number of confirmed cases" and “Number of persons tested negative”, as of May 28, 2020. Until this is clarified, the test positivity in this report is calculated consistently as percent of “number of new confirmed cases” out of “number of new confirmed cases AND number of new persons tested negative.” The number of new cases/persons is a difference between cumulative numbers over two consecutive days, which are published by the state. Then, the test positivity rates are averaged over 7 days.
Note 2: On May 28th, Maryland Health Department also started publishing “Percent positive testing, all jurisdictions”“.
Initially, COVID-19 has affected adult population relatively evenly across different ages (approaching or above 300 per 100,000 population in all age groups 30 and above). Recently, however, the incidence rate has increased more rapidly among those 80 and older (see the second figure). Currently, the incidence rate is highest, 2154.5 per 100,000 population, among people 80+ years of age.
(Source: Maryland Department of Health’s Maryland COVID-19 Case Map Dashboard, and Maryland Department of Planning’s Population Estimates by Race and Hispanic Origin for July 1, 2018)Since testing capacity has been low, we cannot confidently answer this question. Nevertheless, we can see the number of NEW confirmed cases each day. The first figure is for the entire state. We want to see the number of new infections to be stable more or less, without sudden and large increases. On 2020-07-20, 554 new cases were confirmed, compared to 925 on the previous date. The (now relatively small) peak on March 28th was due to new cases in Carroll county on March 28 - see the second figure below.
Hover over each figure to see values and more options.
Now, among 22 counties with 200 or more confirmed cases, below shows the trends of daily number of NEW confirmed cases for each of the county, with 7-day rolling averages in black lines. Again, large spikes on April 8th may well be results of delayed processes in laboratory and/or data entry, not actual increases on that date.
Hover over each figure to see values and more options.
[1] 22Finally, below shows counties ranked by the latest level of daily new cases (i.e., right end of the black line above) - per 100,000 population this time. For example, based on the latest 7-day average, there are 14 new cases per day per 100,000 population in Prince George’s county.
As of 2020-07-20 10:00AM, 78685 confirmed cases have been reported in Maryland, according to Maryland Department of Health. This means there are 1302 people with confirmed COVID-19 per 100,000 population in the state.
Below figure shows number of confirmed cases (gray bars) and infection rates (i.e., number of confirmed cases per 100,000 population) (blue bars) by county, as of 2020-07-20.
Hover over each figure to see values and more options.
Data sources:
1. All COVID-19 data for Maryland come from Maryland COVID-19 Case Map Dashboard published by Maryland Department of Health. This dashboard presents latest numbers on tests, cases, hospitalizations, and deaths as of 10:00AM on each day. Accessed on 2020-07-20. County-level data from New York Times’ Coronavirus in the U.S.: Latest Map and Case Count are no longer used, as of July 19, 2020. 2. All data on county population come from US Census Bureau’s County Population Totals: 2010-2019. Accessed on March 29, 2020.
3. Data on Maryland population by age and sex come from Maryland Department of Planning’s Population Estimates by Race and Hispanic Origin for July 1, 2018. Accessed on April 1, 2020.
See GitHub for data, code, and more information. For typos, errors, and questions, contact me at www.isquared.global.
Making Data Delicious, One Byte at a Time, in good times and bad times.